OK, we did find the bug in the macb driver, the main thread on netdev is here:
https://lore.kernel.org/netdev/02874ece860811409154e81da85fbb5892325...@orsmsx121.amr.corp.intel.com/ For future reference, the main issue is that the driver was doing the timestamping for all socks, so when a sock had NOT requested SOF_TIMESTAMPING_TX_HARDWARE it was getting woken up anyway because the timestamp is available in the error queue. This bug can be closed, I don't know if I need to do anything else? Thanks, Paul -- "This e-mail message and any attachments are confidential and may be privileged. If you are not the intended recipient please notify AMSC immediately by replying to this message or by sending a message to postmas...@amsc.com <mailto:postmas...@amsc.com> and destroy all copies of this message and any attachments. Thank you."